Great way to approach the problem. As someone else mentioned, cheaper launches might adjust the overall risk/cost dynamics but yeah, this data point makes that seem like a stretch.

A big part of the high cost of the project is that they know they only get one and they want to get the most bang per kg of payload.

Most of the cost was research, development and testing of new technologies. Building 10x JWST would not cost 10x what we paid for a single telescope. A single launch also means everything has to be perfect, there is no prototype in space, no second try and no learning from mistakes.

The project would also probably have made drastically different design and management decisions if they knew in advance they were sending >5 units instead of one.

Wouldn’t most of that cost be baked into the launch rather than the payload though?